Skip to content

Consider optional vector search via external embedding APIs #154

@stupidloud

Description

@stupidloud

I saw that local embeddings / vector search were removed because of package size, model downloads, WASM / ONNX memory pressure, and reliability issues.

Would you consider bringing vector search back as an optional feature using external embedding APIs instead of bundling an embedding model inside CodeGraph?

For example, users could configure an OpenAI-compatible embeddings endpoint, Google Gemini Embeddings, or Cloudflare Workers AI embeddings.

Google Gemini API appears to provide a free tier for experimentation, so this may let users try semantic search without CodeGraph shipping or running the embedding model itself.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions